This is a natural history study designed to identify markers of early malignant transformation that could be used in predicting which oral soft tissues are most likely to lead to oral cancer. The scientific questions are: Are these molecular markers that are correlated with more histologically or clinically advanced lesions or predictive of histologically- or clinically-defined progression of oral soft tissue lesions? (Histological gradiants are defined in several dimensions including hyperplasia, metaplasis, dysplasia, carcinoma-in-situ, invasive carcinoma.) Are there behavioral factors and molecular interactions that are correlated with or predictive of histological or clinical status or histological or clinical progression of oral soft tissue lesions? How strong are the associations among the behavioral factors, the most predictive markers, and clinical or histologic status? Persons with white and or red oral mucosal lesions seen at six Veterans Affairs hospitals will be selected for the study. Study subjects will undergo a biopsy of all suspect oral lesions, provide information about tobacco, alcohol intake, and other habits and re- examined every four months for two years and re-biopsied as appropriate. The biopsy material will be analyzed for the presence or activity of certain molecules thought to be associated with the malignant phenotype including p53-dependent growth suppressors, pRb- dependent growth control factors, growth factor receptor-modulated pathways, and viral agents. Statistical analysis will determine associations between molecular markers from the oral lesions and the lesions' histopathology and will examine how environmental, behavioral, and sociodemographic factors influence these associations. If molecules indicative of early transformation can be identified, they may be used to identify persons who should be monitored most closely for the development of oral cancer and to recognize oral cancers at the earliest, most treatable stage.